Prevalence of PCOS and related hyperandrogenic traits in premenopausal women with type 1 diabetes: a systematic review and meta-analysis

Abstract

Abstract BACKGROUND An increased prevalence of functional hyperandrogenism—including polycystic ovary syndrome (PCOS)—has been described in women with type 1 diabetes (T1D). However, heterogeneity between studies is frequent, and prevalence rates vary according to different criteria used for the diagnosis of PCOS and the population studied. OBJECTIVE AND RATIONALE We aimed to perform a systematic review and meta-analysis of the prevalence of PCOS and related hyperandrogenic traits in premenopausal women with T1D. This way, we intend to increase the precision of the estimates of prevalence of PCOS and related traits in women with T1D, and to explore sources of heterogeneity while providing prevalence estimates for clinically relevant subgroups such as the different phenotypes. SEARCH METHODS We conducted a systematic review of the literature using Medline-OVID and Embase databases (Open Science Framework registry for systematic review protocols, https://osf.io/6cv9p/). Studies published up to 29 March 2021 were considered. We selected cross-sectional or prospective studies that reported, in patients with T1D, prevalence data on PCOS according to current definitions and different phenotypes, and/or prevalence rates of other related traits (hirsutism, hyperandrogenaemia, oligo-amenorrhoea and/or polycystic ovarian morphology: PCOM). Exclusion criteria for the review were studies addressing types of diabetes other than T1D; and studies using diagnostic definitions of PCOS different than those mentioned above. Two independent researchers performed data extraction. To assess the risk of bias, we used a tool developed specifically to appraise population-based prevalence studies. OUTCOMES We selected 19 studies (1042 women) reporting the prevalence of PCOS and/or other hyperandrogenic traits. Regarding bias, 12 studies were considered of low-risk, and the remaining seven studies were considered intermediate risk. The pooled prevalence of PCOS when considering all possible phenotypes (ESHRE- American Society for Reproductive Medicine criteria) in T1D was 26% (95% CI: 19–34%; 13 studies, 684 women). Pooled prevalence of classic PCOS (US National Institutes of Health criteria) was 16% (95% CI: 10–22%; 9 studies, 614 women). Pooled prevalence of hyperandrogenic PCOS (Androgen Excess and PCOS Society criteria) was 26% (95% CI: 16–41%; 5 studies, 329 women). Hirsutism (24%), hyperandrogenaemia (29%), oligomenorrhoea (24%) and PCOM (34%) were also prevalent. Heterogeneity was high in almost all these meta-analyses. WIDER IMPLICATIONS This systematic review and meta-analysis showed that PCOS and related hyperandrogenic traits are present in approximately one in every four women with T1D. Larger studies are needed to confirm this association, to address the effect of different variables on the occurrence of PCOS.